Moderated Mediation with JASP

How to test a moderated mediation (conditional indirect effects) with models 7, 8, 14 ,15 (model numbers based on the PROCESS models by A. Hayes) using JASP.
You can find the necessary lavaan-Code for these four models here:
www.regorz-stat...
How to report a moderated mediation:
• How to Report a Modera...

[@user-pp1qf3sz8r]

Thank you so much for this video. It was really helpful!!!

[@georgearthurmantiri]

Thank you very much

[@huisphinx5385]

Thank you for this helpful video!!! However, I encountered a problem. When entering "COMMAND+ENTER", this is an error saying "the INT_AC" can't be recognized". I followed your instruction and changed all the names accordingly.

[@RegorzStatistik]

Do you have in your dataset the variable INT_AC and you still get this error message? If that is the case I don't know the reason for that.

[@huisphinx5385]

@@RegorzStatistik Thank you for your reply. I figured it out. I deleted the long term name then it works. Thank you for the tutorial super helpful :)

[@daphneromero65]

Hi Sir! If I may ask, using JASP, how can we possibly center (i.e., mean centering) the IV and the moderator before conducting moderation analysis? Thank you so much!

[@RegorzStatistik]

You could use the "Create Computed Column" function in JASP. If you search KZbin with the search string , then you will find a couple of videos explaining that. For mean centering you substract the mean of a variable from the variable, e.g. for the IV: IV - mean(IV)

[@daphneromero65]

@@RegorzStatistik Thank you very much Sir!

[@alisonpoor181]

Thanks for this helpful video! Do you know if it is possible to double mean center in JASP? I've seen this as a recommendation for product indicators to do examine the interaction of latent variables. Thanks!

[@RegorzStatistik]

This video is based on path analysis, not on latent variables. I don't know it JASP supports double mean centering.

[@annabeldeclercq3819]

Is it possible to run model 59 in JASP?

[@RegorzStatistik]

I believe that should be possible in principle by using path modeling as in these examples. However, I don't have the syntax code for that.

[@robbieharper3446]

Thanks for the video! Any idea how to code for a moderated mediation using a categorical (base coded) independent variable with three variables? Also, any advice on how to do a power analysis for a moderated mediation?

[@robbieharper3446]

Using model 7 (moderated a path) by the way.

[@RegorzStatistik]

In that case it would be a much more complex path model because you have more than one IV in the model. I don't know of any accessible way to do power analysis for a moderated mediation.

[@Cjstistajh]

I wonder can you analyse moderated moderated-mediation in JASP

[@RegorzStatistik]

In principle yes, I think. You just would to have rebuild all models that e.g. the PROCESS macro uses to test for moderated moderated mediation with the lavaan syntax. However, in this case I think it is easier just to use PROCESS for R (not necessarily for all analyses, but at least for the moderated moderated mediation) - even if one does not uses R on a regular basis.

[@andinamegalarasati4465]

Thank you for the generous tutorial sir. Can I ask the regression syntax for moderated mediation model 7 (no moderation of the direct effect)? Thank you before 🙏

[@RegorzStatistik]

You can find the syntax for the four models (7, 8, 14, 15) on the companion webpage to this tutorial: www.regorz-statistik.de/en/moderated_mediation_JASP.html#model7

[@andinamegalarasati4465]

@@RegorzStatistik Danke schön!

[@sunday_latte]

Is it possible to probe the interaction for Model 8 on JASP? Thank you for this very helpful tutorial!

[@RegorzStatistik]

Yes, you can probe the interaction for the moderated a-path. If you follow the link to the lavaan code for Model 8 in the video discription you will find code for conditional indirect effects. I think if you drop the "* b1"-part of the code, then you will have code for probing the interactions for the a-path.

[@sunday_latte]

@@RegorzStatistik Thank you! I will check this out

[@brostache22]

Is it possible to include control variables in this type of analysis?

[@RegorzStatistik]

Yes, you simply include your control variables in both regression models (for the a-path and for the b/c'-path).